Commit Graph

1332 Commits (211ba4a3dec4ef2457293f33ba8431b6dab6a74e)
 

Author SHA1 Message Date
Alexis Lahouze 7eeb895d3d Rolled back comment on filesystem creation. 2014-07-08 10:15:38 +02:00
Leseb 67ce4907e9 Merge pull request #93 from leseb/wip-package-key
Use a local key for package instead of downloading it
2014-07-03 16:26:50 +02:00
Sébastien Han 9d7c23396f Use a local key for package instead of downloading it
Using a local key is significantly faster than "wgetting" the key from a
server on Internet.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-07-03 15:53:13 +02:00
Leseb cd74f9fe51 Merge pull request #92 from leseb/use-latest-ceph
Use firefly as the default release
2014-07-03 15:37:49 +02:00
Sébastien Han bd03dd9b02 Use firefly as the default release
Since Firefly is out, we use it.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-07-03 15:36:59 +02:00
Leseb a3c1e1c862 Merge pull request #91 from leseb/wip-osd-directory
Ability to use a directory for OSD instead of a disk
2014-07-03 15:34:57 +02:00
Sébastien Han 4f5d195788 Ability to use a directory for OSD instead of a disk
This commit implements a fourth scenario where we can directely use a
directory instead of a block device for the OSDs. The purpose of this
scenario is more testing-oriented. Please note that we do not check
the filesystem underneath the directory so it is really up to you to
configure this properly. Declaring more than one directory on the
same filesystem will confuse Ceph.

Fixes: #14

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-07-03 15:30:49 +02:00
Alexis Lahouze 132e78df84 Allow devices to be partition instead of disks (used for some physical servers with a sigle disk for system and storage). 2014-06-26 14:19:09 +02:00
Alexis Lahouze 8b3f428f1f Replaced all by the list of host groups to be able to run the playbook with non-ceph servers. 2014-06-26 09:52:06 +02:00
Leseb a608b65cdd Merge pull request #86 from maethor/patch-1
Update README.md
2014-06-13 16:36:49 +02:00
Guillaume Subiron 0ec9298b93 Update README.md
Two lines are cut in the README. This patch fix this using http://www.sebastien-han.fr/blog/2014/03/03/Announcing-ceph-ansible-playbooks/.
2014-06-13 16:30:50 +02:00
Leseb 422629557d Merge pull request #85 from kchristensen/fix-architecture-tests
Fix Ansible architecture tasks
2014-06-02 21:06:28 +02:00
Kyle Christensen 168faef3bd Fix Ansible architecture tasks 2014-06-02 15:00:43 -04:00
Leseb ba07e01cb1 Merge pull request #84 from leseb/fix-no-ext4
Since Emperor, Ceph stores xattr into levelDB thus this option is not
2014-05-20 15:14:48 +02:00
Sébastien Han 670bec24c7 Since Emperor, Ceph stores xattr into levelDB thus this option is not
needed anymore.

Fixes: #83

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-05-20 15:04:09 +02:00
Leseb 32b9a84dcd Merge pull request #82 from lorin/master
Use Vagrant's Ansible provisioner
2014-05-12 06:42:28 -07:00
Lorin Hochstein 92c0445989 Use Vagrant's Ansible provisioner
Use Vagrant's built-in support for Ansible provisioner. This eliminates the need
for a hosts file, and simplifies the ansible config file.

Renames config from .ansible.cfg to ansible.cfg since Ansible expects the file
to be called ansible.cfg and be adjacent to Vagrantfile when using the Vagrant
provisioner.
2014-05-11 19:34:37 -04:00
Leseb 3b68622f9d Merge pull request #81 from leseb/enh-testing-mode
Enable single host mode
2014-05-01 01:27:20 +02:00
Sébastien Han 567c0bb6ce Enable single host mode
This commit introduces a new config option 'osd crush chooseleaf type'.
With the help of this option and by setting it to '0' we tell Ceph to
store all the replicas on a single host. Basically we tell CRUSH to
iterate over disk and not over host.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-05-01 01:24:20 +02:00
Leseb d6a82a6201 Merge pull request #80 from leseb/change-default-disk
Change the default disk numbering
2014-04-30 17:58:51 +02:00
Sébastien Han ed52853d20 Change the default disk numbering
While trying to auto-provision with vagrant, new disks get /dev/sdb and
so forth. So starting from /dev/sdd doesn't make sense.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-04-30 17:57:10 +02:00
Leseb 358b58bd4a Merge pull request #79 from leseb/fix-rgw-email
Fix non-declared rgw email_address variable
2014-04-30 14:58:02 +02:00
Sébastien Han f0b9994a03 Fix non-declared rgw email_address variable
Fix: #77

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-04-30 14:56:37 +02:00
Leseb b0d9090cc6 Merge pull request #76 from AcalephStorage/fetch-fsid
Removed dependency on mon[0]
2014-04-30 13:42:08 +02:00
Dexter John Genterone 728e80df5c keys are now copied from the first available mon host
since we're now using fsid for the directory name, it should be safe to
just copy the keys from all mon hosts. Once they are copied, the rest of
the hosts will just skip copying. :)
2014-04-30 16:02:07 +08:00
Hunter Nield 48c1dcd92b Avoid ip/fqdn weirdness & use fsid path for keys 2014-04-30 15:16:11 +08:00
Leseb b4fad6fcfc Merge pull request #74 from guestisp/fix-73-use-ansible-fqdn-in-purge-yml
Fix #73 by using `ansible_fqdn`
2014-04-18 14:01:47 +02:00
Leseb 95d4f4bd1a Merge pull request #72 from guestisp/fix-71-typo-in-purge-yml
Fix #71. Removed wrong `---`
2014-04-18 14:01:36 +02:00
Leseb 2ccbab2c7c Merge pull request #70 from guestisp/fix-69-mon-osd-min-down-reporters
Fix #69 - mon osd min down reporters
2014-04-18 14:01:22 +02:00
Alessandro Corbelli c26b269b5a Fix #73 by using `ansible_fqdn` 2014-04-18 10:55:31 +02:00
Alessandro Corbelli 34706fa613 Fix #71. Removed wrong `---` 2014-04-18 10:52:22 +02:00
Alessandro Corbelli 58ea349b53 Fix #69 - mon osd min down reporters 2014-04-18 10:47:30 +02:00
Leseb 1d5ec9ae1e Merge pull request #68 from leseb/purge-ceph
Add a purge playbook to remove Ceph
2014-04-14 15:19:03 +02:00
Sébastien Han be7f8ab89b Add a purge playbook to remove Ceph
The playbook removes packages, configuration files and ALL THE DATA from
Ceph.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-04-14 15:18:34 +02:00
Leseb 080e2011b4 Merge pull request #67 from leseb/improve-ceph-conf
Improve ceph.conf
2014-04-11 14:08:52 +02:00
Sébastien Han 3c0ffd671f Improve ceph.conf
Add tunables for:

* recovery
* performance

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-04-11 14:07:37 +02:00
Leseb 74eb9eaa04 Merge pull request #66 from leseb/wip-maintenance
Add maintenance playbook
2014-04-09 17:53:55 +02:00
Sébastien Han 9a60f122ef Add maintenance playbook
This playbook was made to automate Ceph servers maintenance
Typical use case: hardware change
By running this playbook you will set the 'noout' flag on your
cluster, which means that OSD **can't** be marked as out
of the CRUSH map, but they will be marked as down.
Basically we tell the cluster to don't move any data since
the operation won't last for too long.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-04-09 17:51:59 +02:00
Leseb 05fdc19182 Merge pull request #65 from guestisp/add-support-for-subdomains-calls
Add support for subdomain calls in S3 api.
2014-04-07 13:29:17 +02:00
Alessandro Corbelli 61ab445128 Add support for subdomain calls in S3 api. See http://ceph.com/docs/master/radosgw/config/#enabling-subdomain-s3-calls 2014-04-07 12:06:24 +02:00
Leseb 6cadebd5c5 Merge pull request #61 from guestisp/remove-default-users-from-rgw
Remove RadosGW default users
2014-03-26 11:35:54 +01:00
Alessandro Corbelli d7195922b2 Totally removed test users creation from RGW tasks 2014-03-26 11:34:05 +01:00
Leseb d2b29bb13b Merge pull request #62 from guestisp/fix-54-use-hostname-in-ceph-conf
Use ansible_hostname in RGW config inside ceph.conf
2014-03-26 11:27:32 +01:00
Alessandro Corbelli c2a0865464 Use ansible_hostname in RGW config inside ceph.conf
Fix #54
2014-03-26 11:26:18 +01:00
Alessandro Corbelli c7476fb375 Remove RadosGW default users 2014-03-26 11:22:37 +01:00
Leseb 4aeaa9609f Update README 2014-03-23 23:20:09 +01:00
Leseb 3ae0b29adb Merge pull request #59 from jcftang/wip-motd-mon
Drop a update-motd.d fragment for monitors so when the user/admin logs i...
2014-03-23 11:59:58 +01:00
Jimmy Tang c3480a588c Drop a update-motd.d fragment for monitors so when the user/admin logs into a monitor there is a small notice for the user about the state of the ceph cluster. Works only on ubuntu for now 2014-03-22 14:57:22 +00:00
Leseb e6c162167c Merge pull request #58 from guestisp/skip-missing-hosts-from-rgw-conf
Skip missing RGW hosts in ceph.conf
2014-03-20 13:42:09 +01:00
Alessandro Corbelli 7623376203 Skip missing RGW hosts in ceph.conf 2014-03-20 12:14:43 +01:00